Instance admins, I've got settings set to WEB_CONCURRENCY=4 and MAX_THREADS=5 on sidekiq. Here's what my dashboard looks like. I'm not sure how to make heads or tails of whether I need more workers or threads. https://cute.group/media/6KQlgwiblNahzJVFBmc
@stickypaws PUMA workers depends on your number of users. Look at your busy states for sidekiq indication.
@stickypaws I see a lot of failures that are just remote hosts being intermittently offline, maybe be sure the failures are actually a problem? I don't see things pilling up in that busy queue so you probably are set for now.
@sharktopus that's good to know
@stickypaws if you click the retries tab up at the top there, what kind of errors are causing the retries? if they're HTTP timeouts it's probably nothing to worry about
@sharktopus 2 pubsubhubbub delivery time outs and 4 remote profile update worker errors that appear to going to a domain that doesn't exist. My gut says in the second case hash.my's server is misconfigured somehow.
@stickypaws That middle graph is just how many tasks that sidekiq has processed over time. Failed numbers are low and your queues don't look busy in that top bar. I'd say you're cool at the moment.
@stickypaws I don't have an answer, but I do note that you're running redis 2.8.17 and I'm running 3.0.6. are there other aspects of your install that might be less than current? just a starting point...
@datn hmmm, I'm using the latest on Debian Jessie, doesn't appear to be a newer version of redis available.
@stickypaws YMMV vary, i can't vouch for them, but dotdeb.org has later versions of redis for Debian. I haven't looked lately, but I feel like redis 3 was a big shift.
@stickypaws Enqueued is 0, so I think you're good.